Build Content Systems That Machines Can Trust
The gap I work in
Most organisations optimise for visibility. SEO for humans, GEO for citations inside AI-generated answers. Both are useful. Neither fixes the part underneath: the publishing systems still produce content that machines cannot reliably read, interpret, or act on.
The result is expensive reconstruction at every read, inconsistent answers across platforms, and content that breaks every time the next platform shift lands.
I fix that. I help organisations modernise their content operations and their publishing pipelines so every output, whether HTML, PDF, DOCX, EPUB, CSV, audio, or video, is machine-ready by design rather than by accident.
MX, the operating system for modern publishing
Machine Experience is the specification layer that makes published content:
- Consistent across every format.
- Self-describing through metadata that travels with the document.
- Deterministic for agents and automation, not something they have to guess at.
- Portable across platforms and time.
- Governed through lifecycle, provenance, and versioning.
Where GEO tweaks prompts and surfaces, MX upgrades the content supply chain itself.
Why organisations need MX-driven publishing
1. Publishing pipelines need structure, not guesswork
Most CMSs and export pipelines produce content that looks fine to humans but is structurally ambiguous to machines. Agents have to reconstruct meaning at read time, a slow, error-prone, compute-heavy process. MX replaces reconstruction with declaration.
2. Metadata is the foundation of machine trust
Every agent begins with the same job: discover context. If identity, provenance, lifecycle, affordances, and semantics are declared, machines can act with confidence. If not, they guess, and guessing does not scale.
3. GEO only works when the content architecture is sound
GEO improves visibility, but it cannot fix broken content structures. MX makes the content usable, so GEO compounds across platforms instead of evaporating with the next vendor change.
What I deliver
MX readiness audits
A full review of your content operations, publishing pipelines, metadata practices, and output formats. I identify where machines fail to read, trust, or reuse your content, and hand back a prioritised remediation plan a content or engineering team can run. Worked argument for the return on this work: why an MX audit pays for itself.
Metadata and governance architecture
A unified metadata layer across all carriers: canonical URLs, version chains, licences, lifecycle, training-data policies, and agent-readiness fields. Verifiable through Reginald notarisation (currently in beta) so downstream consumers, including AI agents, can confirm the content is genuine, unaltered, and authored by who it claims to be.
Publishing system modernisation
Rebuilt export pipelines, semantic HTML, structured DOCX templates, EPUB navigation, CSVW schemas, and machine-readable manifests (`.mx.yaml.md`) at every folder boundary. The cog format underneath is the community-governed standard for documents that machines can read directly, stewarded by The Gathering.
MX-compliant content operations
Workflows, CI gates, and governance rules that ensure every published asset is machine-ready before it leaves the pipeline. Accessibility (WCAG, PDF/UA, the European Accessibility Act), machine-readability, and editorial governance run as one coherent process rather than three.
The outcome
Your organisation gains:
- Reliable machine-readability across every format.
- Lower inference cost for the agents and automation reading you.
- Consistent answers across platforms, not platform-by-platform drift.
- Publishing systems that survive the next shift in the AI vendor stack.
- A single source of truth for identity, provenance, and lifecycle.
The shift is from firefighting content issues to operating at the infrastructure layer of the modern web.
If a deeper read of the philosophy is useful, the MX book series carries the long-form specification: MX: The Intro, MX: The Handbook, and MX: The Protocols.
In one line
I help organisations modernise their content operations so every published asset is trustworthy, structured, and usable by every machine that will ever read it.